How Much Network Lag is too Much?

Remember when the Internet was first making its way into businesses and homes? Users would wait several minutes as pages loaded line by line. This was acceptable back in the legacy days of dial-up Internet services, before the age of mobile devices and everything else we have today.

Now, however, things have drastically changed. Users are still willing to wait for content to load, but won’t pause for a second longer than they have to. Research shows that applications and websites have mere seconds before a user abandons it in search of something faster.

This raises in important question: How much network lag caused by latency is too much?

“47 percent of users expect pages to load in 2 seconds or less.”

User expectations: Page load times
According to recent statistics from KISSmetrics, a survey of e-commerce shoppers showed that every second counts when it comes to user retention. Researchers found that almost half of all participants – 47 percent – expect pages to load in 2 seconds or less. Overall, 40 percent said they would abandon a site and search for something else if content doesn’t load within 3 seconds. Even a single moment can make a difference – a 1 second delay reduces user satisfaction by 16 percent.

Although these statistics pertain to online consumers, they illustrates the shift taking place with users across nearly every industry. In the current corporate environment, business end users expect a similar – if not higher – level of performance from the mission-critical applications they count on to do their jobs. Just like online shoppers, enterprise employees aren’t willing to put up with a resource that is impacted by more than a few seconds of lag. Reliable, high-speed performance is key. When a company-approved application or technological asset isn’t moving as fast as users would like, it could cause them to turn to unapproved, shadow IT processes.

How does latency impact critical applications?
Network lag can have a considerable impact on critical applications – in some cases, network latency could be so bad that it stunts the platform’s performance to the point that it’s nearly unusable.

“If excessive network latency is causing the application to spend a large amount of time waiting for responses from a distant data center, then the bandwidth may not be fully utilized, and performance will suffer,” noted an except from “IP Storage Networking: Straight to the Core” by Gary Orenstein.

Latency is made up of certain types of delay, including propagation, node, congestion and processing delay. A robust network design can help reduce these types of delays, and minimize overall network lag.

Today's enterprise users won't wait longer than a few seconds for applications to respond.
Today’s enterprise users won’t wait longer than a few seconds for applications to respond.

By the numbers: How much latency is unacceptable?
There are specific factors that can impact the latency of a modem, including the latency of the connecting device, as well as the distance data is traveling.

For a typical modem, the connecting device latency can fall between 5 and 40 milliseconds – a traditional T1 line will normally experience 0 to 10 milliseconds in latency at the connection. It’s also important to factor in distance – data roughly travels at 120,000 miles per second, adding approximately 1 millisecond of delay for every 60 miles traveled. In this way, network lag is not only caused by the connection, but how far data traverses to the data center location and to the receiver.

While these are rough estimates, the numbers are revealing. Any level of connecting device latency or distance latency over these levels could indicate a problem. Similarly, a 2 percent packet loss or more could also be a symptom of performance issues caused by latency.

What can enterprises do about latency?
However, in order to fully understand if the latency or packet loss an enterprise experiences is within normal limits, the IT team must have a way to measure these levels. An SD-WAN solution can provide just the type of in-depth look needed here.

This technology has the ability to track not only latency and packet loss, but jitter and overall availability as well, helping to ensure that each packet takes the most optimal path across the network. An SD-WAN component can create an intelligent network that proactively adapts to current conditions as well as any changes in real time, offering enterprises the speed, performance and reliability they need.

Download this report from Broadband-Testing to see how an SD-WAN improves application performance when compared to a traditional WAN.

Categories: Software Defined WAN (SD-WAN), Application Performance/Application Quality, IT Challenges, Network Reliability